Displacement of Ammonium from Aerosol Particles by Uptake of Triethylamine

Abstract

The displacement of ammonium by triethylammonium (TEAH) in aerosol particles of about 15–35 μm in diameter was investigated using an electrodynamic balance (EDB) coupled with in situ Raman spectroscopy. The phase state of particles played a crucial role in the extent of triethylamine (TEA) uptake. At 50 or 75% relative humidity (RH), the heterogenous uptake of about 40-ppm TEA by aqueous ammonium salts of sulfate [(NH₄)₂SO₄], bisulfate (NH₄HSO₄), nitrate (NH₄NO₃), chloride (NH₄Cl), and oxalate [(NH₄)₂C₂O₄] led to increases in particle mass of over 90%. A complete displacement of ammonium by TEAH was confirmed by direct EDB mass measurements and the Raman spectra obtained. TEAH sulfate was formed during the exposure of aqueous droplets of (NH₄)₂SO₄ and NH₄HSO₄ to TEA vapor at 50% RH; but a fraction of it decomposed to TEAH bisulfate when the TEA supply was removed. Crystalline solid particles of (NH₄)₂SO₄ and (NH₄)₂C₂O₄ experienced small mass increases of <5%, both of which were attributed to the hindered mass transfer of TEA in crystalline solids. However, TEA reacted with the amorphous solid NH₄NO₃ particle at <3% RH as effectively as if it was in the aqueous NH₄NO₃ droplet (50% RH) and formed TEAH nitrate. On the other hand, the amorphous NH₄HSO₄ solid particle reacted with TEA at <3% RH to form crystalline (NH₄)₂SO₄ and liquid TEAH bisulfate and sulfate. The formation of rather inert crystalline (NH₄)₂SO₄ suppressed the ammonium exchange.

Copyright 2012 American Association for Aerosol Research

INTRODUCTION

Short-chain aliphatic amines are emitted from a wide range of sources such as animal husbandry (Schade and Crutzen 1995), vehicle exhausts (Cadle and Mulawa 1980), waste incinerators and sewage treatment plants (Leach et al. 1999). In intense agricultural regions, ambient levels of amines could reach hundreds of parts per billion (ppb) (Rabaud et al. 2003), but they are generally lower in other areas. The typical concentrations of amines are <1–100 parts per trillion (ppt) (Gronberg et al. 1992), 1–3 orders of magnitude lower than those of ammonia (Reis et al. 2009). Particulate amines have been detected in fine aerosols (Zhang et al. 2002; Facchini et al. 2008; Sorooshian et al. 2008; Pratt et al. 2009), aqueous fog and rain drops (Collett et al. 2008; Rehbein et al. 2011); and their abundance has been found to be size dependent (VandenBoer et al. 2011). Elevated levels of particle-phase amines were observed during periods of acidic cloud/fog processing (Rehbein et al. 2011), indicating the possible importance of aqueous-phase chemistry and aerosol acidity in particulate amine formation.

Similar to other reactive organic vapors (Kroll and Seinfeld 2008; Hallquist et al. 2009; Li et al. 2011), amines may undergo rapid oxidation with ozone, hydroxyl and nitrate radicals (Angelino et al. 2001; Murphy et al. 2007; Silva et al. 2008; Zahardis et al. 2008) to produce secondary organic aerosols. Moreover, as one of the few atmospheric alkaline vapors other than ammonia (Swartz et al. 1999), amines exhibit acid–base chemistry which is of concern in aerosol science (Ge et al. 2011). Theoretical calculations (Barsanti et al. 2009) and field measurements (Makela et al. 2001; Smith et al. 2010; Wang et al. 2010b) consistently showed that amines formed aminium salts with sulfuric acid, nitric acid, and organic acids, and that they comprised a significant fraction of ambient aerosols, particularly in nucleation mode. The molar ratio of aminium to ammonium in fine aerosol has been reported to range from 0.005 to 0.2 (VandenBoer et al. 2011). Aminium salt formation leads to tremendous decreases in the vapor pressures of the involved species (amines and acids), which may assist nucleation by counteracting the Kelvin curvature effect (Zhang and Wexler 2002; Barsanti et al. 2009). There is evidence that amines enhanced nucleation in sulfuric acid-water system more effectively than ammonia (Kurten et al. 2008). Bzdek, Johnston, and coworkers conducted an extensive investigation on the uptake of amines by small clusters (1–2 nm) of ammonium sulfate, bisulfate, and nitrate using Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR-MS) (Bzdek et al. 2010a; Bzdek et al. 2010b). Complete displacement of ammonium by aminium was observed, suggesting that the sub-3-nm clusters were likely to be aminium salts.

Using reaction flow tubes, the uptake coefficients of low molecular-weight alkylamines (Wang et al. 2010a; Yin et al. 2011) residing on highly acidic sulfuric acid surfaces were measured at various temperatures and sulfuric acid concentrations. Irreversible reactive uptakes were observed. To explore the displacement of ammonium by aminium ion, Qiu et al. (2011) investigated the heterogenous interaction between amines and ammonium sulfate or bisulfate using a coated-wall flow reactor coupled to an ion-drift chemical ionization mass spectrometer (ID-CIMS). Simultaneous loss of amines and generation of ammonia in the gas phase were observed and the results suggested that amines were important contributors to the growth of ambient aerosols. Using a flow tube coupled with ion trap time-of-flight (IT-TOF) aerosol mass spectrometer, Lloyd et al. (2009) investigated the uptake of trimethylamine (TMA) by ammonium nitrate and showed the displacement of NH⁺ ₄ ions by aminium ions.

In this paper, we systematically investigated the heterogenous uptake of triethylamine (TEA) by common inorganic and organic particles. Five salts of ammonium and two salts of sodium in both droplet and solid (crystalline and amorphous) forms under <3, 50, or 75% relative humidity (RH) were studied. An electrodynamic balance (EDB) coupled with Raman spectroscopy (Lee et al. 2008) was employed for simultaneous measurements of the temporal changes in mass, chemical composition, and post-reaction hygroscopicity of levitated droplets and solid particles. TEA [(C₂H₅)₃N] uptake can be tremendously enhanced in an aqueous acidic medium. Protonation of TEA leads to the formation of triethylammonium ions ((C₂H₅)₃NH⁺, abbreviated as TEAH ions hereafter), and shifts the equilibrium to the aqueous phase and further promotes TEA uptake (Ge et al. 2011):

Ammonium (NH⁺ ₄) salts are mildly acidic:

This acidic condition facilitated TEA uptake. In this study, all ammonium-containing aqueous droplets had overwhelming TEA uptake. Particle mass increases were attributed to the formation of TEAH. Complete displacement of ammonium by TEAH was observed for aqueous droplets of ammonium salts and reactions were confirmed directly by matching the product spectra with the Raman spectra of standard compounds collected using the EDB/Raman system.

Aqueous ammonia [NH_3(aq)] was formed after reaction (Equation (3)) and partitioned into the gas phase in the ammonia-free EDB chamber:

EXPERIMENTAL SETUP

1. Single Particle Generation

Ammonium sulfate (>99.0%, Sigma-Aldrich), ammonium bisulfate (>99.5%, Fluka), ammonium oxalatemonohydrate (>99.5%, Fluka), ammonium nitrate (99.999%, Aldrich), ammonium chloride (99.99%, Aldrich), sodium sulfate (>99.0%, Sigma-Aldrich), sodium chloride (>99.5%, Sigma-Aldrich), and the mixture of sulfuric acid (95–98 wt%, Mallinckrodt Chemicals) and TEA (>99%, Aldrich) were dissolved in high-purity Milli-Q water (18.2 Ω) in 10-ml beakers. A small amount of the aqueous solution was introduced to a piezoelectric particle generator (MicroFab Tech., Inc.) and a single droplet of 15–35 μm in diameter (after solvent evaporation) was produced by applying an electric pulse to the generator. The droplet was charged as it passed through the induction plate mounted at the top entrance of the EDB.

FIG. 1 Kinetics of the uptake of 40-ppm TEA by particles of ammonium oxalate, ammonium bisulfate, ammonium sulfate, ammonium chloride, ammonium nitrate, sodium chloride, and sodium sulfate. (Color figure available online.)

2. EDB Equilibrium Mass and Hygroscopic Measurements

EDB has been used extensively in direct measurements of particle mass changes (Davis 1997; Chan et al. 2010). A study on the reactive uptake of nonanal by sulfuric acid/organic mixture has been reported (Chan and Chan 2011). In brief, a charged particle of inorganic or organic salt was trapped and levitated at the null point of the EDB by setting up a combination of AC and DC electric fields around the particle. Assuming that there was no loss of charge, the weight of the single particle would be proportional to the balancing DC voltage. The relative mass change of a particle during TEA vapor uptake can be determined by monitoring the change in the balancing DC voltage. The uptake data obtained was used to calculate the mass ratio, defined as m_t /m_o , and the percentage mass increase, defined as (1 – m_t /m_o ) × 100%, where m_t is the mass of the particle after TEA uptake and m_o is the initial mass of the particle (reference state). For hygroscopic measurements, the mass ratio was defined as the ratio of mass at an elevated RH to the mass at <3% RH.

3. Generation and Detection of TEA Vapor

TEA vapor was generated by evaporating liquid TEA using a syringe pump (kdScientific, KDS-100). The gas-phase TEA concentration was controlled by diluting the TEA vapor-laden air with a stream of purified air at controlled RH using mass flow controllers (Smart-Trak Series 100). The diluted TEA vapor was split into two streams: one for EDB experiments, which was fixed at ∼180 cm³ min⁻¹, and the other was for TEA monitoring with a portable volatile organic compound (VOC) detector (Ion-science Inc., PhoCheck 5000+) and for RH monitoring with a dew point monitor (EdgeTech Inc., Dew Prime I). For all uptake experiments, the EDB stream was allowed to flow once the monitoring stream had become stable and the TEA vapor concentration had a fluctuation of approximately ±5%. Calibration of the VOC detector was done by evaporating known amounts of TEA at controlled flow rates prior to EDB measurements.

4. Raman Spectroscopy of Single Levitated Particles

To probe the temporal chemical changes of levitated particles, a Raman spectroscopy system similar to that of Lee et al. (2008) was employed. The system consisted of a 150-mW diode-pumped solid-state laser (Spectra-Physics Excelsior, 532-150-SLM-CDRH) and a 0.5-m monochromator (Acton SpectraPro 500) attached to a charge-coupled device (CCD) detector (Andor Technology DV420-OE), which was integrated into the EDB system. The 532.2-nm line of the laser was used as the source of excitation. A pair of lenses, which matched the f/7 optics of the monochromator, was used to focus the 90° scattering of the levitated droplet in the EDB onto the slit of the monochromator. A Raman notch filter was placed between the two lenses to remove the strong Rayleigh scattering. A 300 g mm^–1 grating was selected for the monochromator. The integration time of each spectrum was 10 s (10 frames, each with an accumulation time of 1 s). The resolution of the spectra obtained was about 6 cm⁻¹. The assignments of Raman peaks were based on those used by Dawson et al. (1986), Lin-Vien et al. (1991), and Dong et al. (2007).

RESULTS AND DISCUSSION

1. Enhanced TEA Uptake of Ammonium Salts

Levitated droplets and solid particles of the chosen inorganic and organic salts were exposed to 40-ppm TEA in an EDB for 24 h at controlled RHs (<3, 50, or 75%). Their uptake kinetics are shown in Figure 1.

The droplets of neutral salts, such as sodium chloride and sodium sulfate (NaCl and Na₂SO₄, 75% RH), had negligible mass changes during the course of TEA exposure. Using the Henry's Law constant of TEA in water (6.6 M atm^–1), the estimated mass uptake of the aqueous droplets equilibrated at 40-ppm TEA vapor was <0.1%, consistent with the observed constant NaCl and Na₂SO₄ masses.

FIG. 2 (a) Temporal changes in Raman signals of an aqueous ammonium chloride droplet at 40-ppm TEA and (b) comparison of the Raman spectra of a solid particle of ammonium chloride, the dried particle of the reacted ammonium chloride droplet in TEA-free EDB chamber, and a solid particle of TEAHCl. (Color figure available online.)

TEA uptake can be dramatically enhanced in an aqueous acidic medium. Protonation of TEA leads to the formation of TEAH ions and shifts the equilibrium to the aqueous phase [Equations (1) and (2)]. Ammonium salts are mildly acidic [Equation (3)] and their aqueous solutions facilitated TEA uptake. In this study, all ammonium-containing droplets had overwhelming TEA uptake. Particle mass increases were attributed to the formation of TEAH and will be discussed in the detailed Raman analysis later. At 50 or 75% RH, aqueous droplets of ammonium oxalate [(NH₄)₂C₂O₄], ammonium bisulfate (NH₄HSO₄), ammonium sulfate [(NH₄)₂SO₄], ammonium chloride (NH₄Cl), and ammonium nitrate (NH₄NO₃) had mass increases of more than 90% after 1 h of exposure to 40-ppm TEA. However, the reactions of TEA with ammonium crystalline solids were inefficient. Crystalline solid particles of (NH₄)₂SO₄ and (NH₄)₂C₂O₄ had mass increases of less than 5% after 24 h of exposure. The phase states of particles played an important role in TEA uptake. To elucidate the changes in particle mass, detailed chemical analysis was carried using Raman spectroscopy. The analysis of aqueous ammonium salts is discussed first and then followed by the discussion of crystalline and amorphous solid ammonium salts.

TABLE 1 Summary of the reaction conditions, the particle mass changes on a dry basis and the theoretical mass ratios

Particle species	Relative humidity (%)	Physical state	^a Exposed/initial particle mass (dry basis)	^b Theoretical mass ratio
NH4Cl	50	aqueous	2.59	2.57
NH4NO3	50	aqueous	2.07	2.05
NH4HSO4	50	aqueous	1.86	1.84
(NH4)2SO4	50	aqueous	1.57	1.60
(NH4)2SO4	75	aqueous	1.59	1.60
(NH4)2SO4	<3	crystalline	1.03	N/A
(NH4)2C2O4	<3	crystalline	1.05	N/A
NH4NO3	<3	amorphous	2.01	2.05
NH4HSO4	<3	amorphous	1.43	1.44
^a The EDB-measured mass ratio between the reacted particles subsequently dried at <3% RH in the absence of TEA vapor (the exposed particle dry mass) and the initial particle dry mass.
^b Calculated molecular mass ratio between the expected product (e.g., TEAHCl) and the initial reactant (e.g., NH4Cl).

2. Raman Analysis of NH₄ ⁺ Displacement and TEAH Formation in Aqueous NH₄ ⁺ Salts

Temporal chemical changes of an aqueous NH₄Cl droplet when reacting with 40-ppm TEA at 50% RH were monitored by Raman spectroscopy (Figure 2a). The spectrum of the aqueous NH₄Cl droplet at 50% is shown as a control.

At 10 min, the totally symmetric stretching mode of NH₄ ⁺ at 1662 cm⁻¹ vanished. On the other hand, the symmetric stretching mode of carbon—carbon–nitrogen (C–C–N) at 899 cm⁻¹ and the unresolved multiple peaks of combined CH₂/CH₃ stretching at 2800–3200 cm⁻¹ emerged due to the formation of TEAH ions. The substantial formation of TEAH suggested that protonation of TEA was effective, resulting in a huge mass increase (110%, Figure 1). Concurrently, the removal of proton by TEA promoted the dissociation of ammonium into aqueous ammonia (Equation (3)). Aqueous ammonia eventually partitioned into the gas phase in the ammonia-free EDB chamber (Equation (4)), causing the ammonium signals to disappear. The EDB chamber was connected to a continuous flow system and the ammonia generated by the single levitated particle was not expected to accumulate throughout the experiment. Overall, ammonium ions were displaced by TEAH ions in the aqueous NH₄Cl droplet. After the initial 10 min, the Raman features of the reacted NH₄Cl droplet remained constant, consistent with the observation of a stable particle mass (Figure 1). The ultimate spectrum of the reacted NH₄Cl droplet at 24 h was similar to the standard spectrum of triethylammonium chloride (TEAHCl) droplet, confirming the displacement of ammonium by TEAH. The reacted NH₄Cl droplet was dried at <3% RH in the absence of TEA vapor and its spectrum is compared with the spectra of unexposed pure NH₄Cl particle and pure TEAHCl particle collected using the EDB/Raman system (Figure 2b). Again, the spectra of the reacted NH₄Cl particle matched most of the TEAHCl Raman features.

Table 1 lists the mass ratios of the reacted particles (<3, 50, or 70% RH; 40-ppm TEA) subsequently dried (<3% RH) in the TEA-free EDB chamber to the corresponding initial particles and the theoretical mass ratios (see also the footnotes of Table 1).

The EDB-measured dry mass ratio of the reacted particle to the initial NH₄Cl solid particle was very close to the molecular weight ratio of TEAHCl to NH₄Cl (theoretical mass ratio), validating the use of the coupled EDB/Raman system in ammonium/triethylammonium exchange studies. Similar to the NH₄Cl droplet, the NH₄NO₃ droplet reacted with TEA to form TEAH nitrate at 50% RH (Supp. Info., Figure S1a). The ammonium displacement was confirmed by matching the Raman signals of the exposed NH₄NO₃ droplet with the signals of nitrate and TEAH from NH₄NO₃ and TEAHCl droplets, respectively. The generalized overall reaction of aqueous (mono-) ammonium salts was:

where X represents Cl⁻ or NO₃ ⁻.

3. Decomposition of TEAH Sulfate into TEAH Bisulfate

As will be shown later, reactions of TEA with aqueous (50 or 75% RH) (NH₄)₂SO₄ and NH₄HSO₄ particles form TEAH sulfate. However, TEAH sulfate can convert to TEAH bisulfate (and vice versa) and hence a discussion of the conversion is given here before the data analysis of the uptake by aqueous (NH₄)₂SO₄ and NH₄HSO₄ droplets. We examined the uptake of TEA by TEAH bisulfate (TEAHBS) droplet, which was synthesized by mixing TEA with diluted aqueous sulfuric acid solution in a 1:1 molar ratio, for Raman characterization. From the observed Mie scattering patterns, TEAHCl particles crystallized at <3% RH while the mixed particle of TEA and sulfuric acid (i.e., TEAHBS) did not and remained in droplet form. Liquid TEAHBS is commonly referred to as a protic ionic liquid (Drummond and Greaves 2008) and has a reported melting point of 84.2°C (Belieres and Angell 2007). Interestingly, the single levitated TEAHBS droplet remained as a supercooled liquid (Debenedetti and Stillinger 2001; Zobrist et al. 2008) and did not crystallize at room temperature (∼23°C), about 61°C below its melting point. Suppressed crystallization has often been observed in hygroscopic studies involving an EDB (Peng et al. 2001). To show the presence of TEAH ions in the synthesized TEAHBS, the Raman spectrum of the synthesized droplet was compared with the spectra of NH₄HSO₄ and TEAHCl particles at <3% RH (Supp. Info., Figure S2a). TEAH ions were identified in the synthesized TEAHBS particles, as confirmed by the presence of characteristic CH₃ deformation at 1450 cm⁻¹ (shaded), unresolved multiple peaks of CH₂/CH₃ stretching at 2800–3200 cm⁻¹ and C–C–N stretching at ∼899 cm⁻¹. Sulfate ions (SO₄ ²⁻) are characterized by peaks at 452, 613, 973, and 1080 cm⁻¹ from (NH₄)₂SO₄ while bisulfate ions (HSO₄ ⁻) are characterized by peaks at 411, 428, 583, 600, 870, and 1013 cm⁻¹ from NH₄HSO₄. All bisulfate Raman features were identified in the spectrum of TEAHBS particle but sulfate peaks were absent. Overall, the Raman features of the TEAHBS particle were the superposition of the peaks of TEAH and bisulfate ions, confirming that the mixing of TEA and sulfuric acidin a 1:1 molar ratio produced TEAHBS. In aqueous state (50% RH, Supp. Info., Figure S2b), NH₄HSO₄ dissociated into sulfate and bisulfate ions (Clegg et al. 1998). Again, the Raman features of aqueous TEAHBS droplet were the superposition of the TEAH and bisulfate features of TEAHCl and NH₄HSO₄ droplets, respectively. No sulfate peaks were observed in the spectrum of TEAHBS droplet.

Triethylammonium sulfate (TEAH sulfate) was produced when particles of TEAHBS were exposed to the 40-ppm TEA vapor. The uptake was studied at 50 and <3% RH. A high RH favored the formation of TEAH sulfate. At 50% RH (Figure 3), the TEAHBS particle had a mass increase of about 80% while the reaction at <3% RH yielded a 10% mass increase only. Most of the uptake occurred in the first 10 min of the reaction.

FIG. 3 Kinetics of the uptake of 40-ppm TEA by TEAHBS droplets at 50 and <3% RH. (Color figure available online.)

FIG. 4 Temporal changes in (a) Raman signals of the aqueous TEAHBS droplet at 40-ppm TEA (50% RH); (b) Raman signals of the aqueous TEAH sulfate droplet during decomposition in the TEA-free EDB chamber (50% RH); (c) Raman signals of the partially decomposed TEAH sulfate particle during decomposition at <3% RH and its Raman spectrum after re-humidification. (Color figure available online.)

The Raman features of the TEAHBS droplet reacted at 50% RH are shown in Figure 4a. At 10 min, the characteristic peaks of bisulfate (411, 428, 589, 886, and 1039 cm⁻¹) reduced significantly and the peaks of sulfates (452, 618, and 973 cm⁻¹) emerged while the TEAH Raman features persisted, confirming that TEAH sulfate was indeed formed. Except for the potential overlap with the bisulfate peak at 1039 cm⁻¹, the characteristic peaks of TEAH ions in TEAH sulfate at 1000–1100 cm⁻¹ and the three peaks of TEAHCl at the same Raman shift were the same. No spectroscopic change was observed beyond 10 min, which was consistent with the observed stable mass of reacting TEAHBS droplet during the same period (Figure 3).

TEAH sulfate was stable in the presence of gaseous TEA. After 24 h of exposure, TEA supply was cut off and replaced by TEA-free air at the same 50% RH. In Figure 4b, the temporal spectral changes of the TEAH sulfate droplet in the absence of TEA vapor are shown. Sulfate signals weakened whilst bisulfate signals strengthened, indicating that TEAH sulfate decomposed into aqueous TEAH and TEAHBS, albeit at a much lower rate than the initial TEA uptake. A mass decrease of about 25% was measured with the EDB 7 h after the TEA supply was cut off, suggesting that a significant amount of particle-phase free-TEA, from both decomposition and absorption, escaped into the gas phase. Since the RH was kept constant at 50% during the decomposition, the contribution of water evaporation to the mass decrease was expected to be less significant compared with the liberation of free-TEA. Afterwards, the particle was left in the EDB chamber for another 14 h for drying at <3% RH. In the first hour, the sulfate peak at 973 cm⁻¹ was significantly weaker while the bisulfate peak at 1039 cm⁻¹ was stronger than those before the particle was dried (but after it was exposed to TEA-free air at 50% RH) (Figure 4c, <3% RH).

The particle then became highly fluorescent after being dried for 14 h, masking the Raman features. The same particle was then rehumidified at 50% RH and the Raman features of sulfate and bisulfate were found to be similar to those of the particle dried for 1 h only. This suggested that the conversion of sulfate to bisulfate may have ceased after 1 h at <3% RH. It is noted that from our later experiments, trace amounts of gaseous TEA may have remained in the EDB chamber which might have then suppressed further decomposition of TEAH sulfate to liberate TEA. The percentage of TEAH sulfate remained after drying was estimated using the EDB-measured dry mass of the TEA-exposed particle at 14 h and that of the initial TEAHBS. The molar ratio of TEAHBS to TEAH sulfate was determined to be approximately 88:12. This ratio was used as the ratio of stable TEAH sulfate to TEAHBS after drying at <3% RH in a later analysis of uptake experiments of aqueous (NH₄)₂SO₄ and NH₄HSO₄ droplets.

In additional experiments that the TEAHBS droplet reacted with TEA vapor at <3% RH, the sulfate signal at 973 cm⁻¹ increased while the bisulfate signal decreased but did not vanish (Figure s3a, Supp. Info). This was different from the reaction at 50% RH at which the bisulfate signal had almost disappeared and the conversion to TEAH sulfate was complete (Figure 4a). In the TEA uptake experiment at <3% RH, only a portion of TEAHBS was converted to TEAH sulfate, leaving a mixture of TEAH sulfate and TEAHBS at 24 h. As shown in the uptake kinetics (Figure 3), the mass increase of the TEAHBS particle reacted at <3% RH was 8 times smaller than that reacted at 50% RH. We speculate that TEAH sulfate may not be stable in highly concentrated droplet at <3% RH. Twenty hours after cutting off the TEA supply (Supp. Info., Figure S3b), TEAH sulfate decomposed and the mixed TEAH sulfate/bisulfate particle (formed at <3% RH, 40-ppm TEA) transformed into a TEAHBS particle with Raman features resembling those of the initial TEAHBS particle.

FIG. 5 (a) Temporal changes in the Raman signals of an amorphous ammonium bisulfate solid particle at 40-ppm TEA (<3% RH) and (b) comparison of the Raman spectra of the dried particle of the reacted ammonium bisulfate in TEA-free EDB chamber, a 50 mol% ammonium sulfate/TEAHBS mixed particle, a crystalline ammonium sulfate solid particle, and a TEAHBS droplet, all at <3% RH. (Color figure available online.)

4. Reactions of Aqueous NH₄HSO₄ and (NH₄)₂SO₄ Droplets with TEA

As indicated from the changes of Raman spectra, both droplets of NH₄HSO₄ and (NH₄)₂SO₄ reacted with TEA to form the same reaction product, TEAH sulfate, at 50 or 75% RH. Ammonium ions were replaced by TEAH ions as reflected by the loss of NH₄ ⁺ peaks and the advent of C–C–N stretching due to TEAH (Supp. Info., Figures S4a–c). Unlike aqueous (NH₄)₂SO₄ droplets and other aqueous droplets of ammonium salts used in this study (NH₄Cl, NH₄NO₃),aqueous NH₄HSO₄ droplets provided extra protons from the proton dissociation of bisulfate ions for the neutralization reaction with TEA to form TEAH. Both neutralization and ammonium displacement were responsible for the mass and chemical changes of the aqueous NH₄HSO₄ droplet. The overall reaction of aqueous NH₄HSO₄ droplet with TEA led to the removal of one mole of ammonium per mole of NH₄HSO₄. Stoichiometrically, it can be represented by:

On the other hand, aqueous (NH₄)₂SO₄ (at both 50 and 75% RH) droplets reacted with TEA and two moles of ammonium were removed per mole of (NH₄)₂SO₄:

The TEA uptake and ammonium displacement were effective in aqueous droplets. The mass increase of aqueous NH₄HSO₄ droplet after 40-ppm TEA uptake at 50% RH was about 160% (with reference to aqueous NH₄HSO₄ at 50% RH). (NH₄)₂SO₄ droplets reacted at 50 and 75% RH had mass increases of about 130%. One extra ammonium was removed from the (NH₄)₂SO₄-TEA reaction compared with the NH₄HSO₄-TEA reaction but the same reaction product, TEAH sulfate, was formed. Hence, the mass increases of (NH₄)₂SO₄ droplets after a 24-h exposure were less than that of the NH₄HSO₄ droplet in aqueous state. It is noted that the mass increases during TEA exposure (Figure 1) were complicated by the difference in hygroscopicity between parent [(NH₄)₂SO₄ or NH₄HSO₄] and reacted (TEAH sulfate) droplets, and the potential mass contribution by the physical partitioning of TEA vapor. Nevertheless, the EDB-measured dry mass ratios of exposed to initial particles of NH₄HSO₄ and (NH₄)₂SO₄ were consistent with the theoretical mass (Table 1). A mass ratio of TEAHBS to TEAH sulfate of 88:12 was assumed after decomposing the reacted NH₄HSO₄ and (NH₄)₂SO₄ droplets (i.e., TEAH sulfate) at <3% RH in the TEA-free EDB chamber, as discussed in Section 3.

5. Enhanced TEA Uptake by Amorphous Solids over Crystalline Solids

Crystalline solid particles of ammonium oxalate [(NH₄)₂C₂O₄] and (NH₄)₂SO₄ (<3% RH) reacted slowly with TEA, presumably due to the diffusion limitation of the TEA in the crystalline solids. Weak signals of combined CH₂/CH₃ stretching vibrations appeared at 2800–3200 cm⁻¹ in the Raman spectrum of reacted (NH₄)₂C₂O₄ and (NH₄)₂SO₄ particles (Supp. Info., Figure S5), accompanied by a small mass increase of less than 5% after 24 h of reactions (Figure 1).

Unlike (NH₄)₂SO₄ and (NH₄)₂C₂O₄ which formed crystalline solids at <3% RH, NH₄NO₃ particle did not effloresce and may form amorphous solid (Martin 2000). The formation of amorphous NH₄NO₃ solid particle was inferred from the observed light scattering pattern in our experiments that resembled patterns closer to those of droplets than the irregular pattern of crystalline solids. The TEA uptake by amorphous NH₄NO₃ solid particle was as effective as that of the aqueous NH₄NO₃ droplet. In Figure 1, both the NH₄NO₃ droplet (50% RH) and the NH₄NO₃ amorphous solid (<3% RH) had similar equilibrium mass increases of about 97–99% at 24 h. Martin (2000) reported that NH₄NO₃ did not effloresce even under extremely dry condition (∼0% RH) and particle-phase water might have remained in the amorphous solids of NH₄NO₃ at <3% RH. The presence of residual water may have facilitated the ammonium displacement of NH₄NO₃ during reaction with TEA at <3% RH. It is noted that the spectrum of NH₄NO₃ particle that reacted at <3% RH had a doublet at about 450 cm⁻¹ (Supp. Info., Figure S1b) which belonged to neither nitrate nor TEAH. The doublet was also observed in the spectrum when the reacted NH₄NO₃ droplet (50% RH) was dried at <3% RH (Supp. Info., Figure S1c). The origin of the doublet is not known at this point.

Amorphous solid NH₄HSO₄ (Martin 2000; Mikhailov et al. 2009) at <3% RH responded to TEA vapor swiftly and attained a mass increase of about 50% at equilibrium (Figure 1). At 10 min (Figure 5a), the Raman spectrum of reacted NH₄HSO₄ particle contained all features of (NH₄)₂SO₄, including both sulfate and ammonium, and the Raman signals of TEAHBS and TEAH sulfate. Hence, the reaction products were ionic mixtures of TEAH, ammonium, bisulfate, and sulfate.

The physical state (e.g., crystalline or liquid) of reaction products can be identified by analyzing the full width at half maximum (FWHM) (Ling and Chan 2008), also known as peak broadness, of the peaks of interest. In general, Raman peaks of crystalline compounds had smaller FWHM values (or were sharper) compared with their liquid (or aqueous) counterparts. The sulfate peak at 613 cm⁻¹ was investigated. The FWHM(613) of liquid TEAH sulfate was 39.3 cm⁻¹, almost double that of crystalline (NH₄)₂SO₄ (20.8 cm⁻¹). The FWHM(613) of reacted NH₄HSO₄ was 21.2 cm⁻¹, suggesting that the sulfate of reacted NH₄HSO₄ particle was in crystalline state. As deduced from the list of product ions mentioned above, the crystalline phase is likely (NH₄)₂SO₄. It should be noted that TEAHBS partially reacted with TEA at <3% RH, forming a fraction of TEAH sulfate. Hence, in addition to that of crystalline (NH₄)₂SO₄, mixed signals of TEAH sulfate and TEAHBS are expected if TEAHBS is produced from NH₄HSO₄-TEA reactions. In our experiments at<3% RH, TEAHBS dominated over TEAH sulfate. When Raman signals of TEAHBS were subtracted from the spectrum of reacted NH₄HSO₄ particle, the resultant spectrum matched most features of crystalline (NH₄)₂SO₄ particle, except for the potential contribution of sulfate signals from liquid TEAH sulfate. The reacted NH₄HSO₄ particle showed the Mie scattering pattern of a mixed-phase liquid/solid particle (Figure 5a), consistent with the deduction of crystalline (NH₄)₂SO₄ mixed in liquid TEAHBS and TEAH sulfate mixtures. Overall, the reaction of NH₄HSO₄ at <3% RH is proposed to be:

This is in contrast to the reaction of aqueous NH₄HSO₄ droplet with TEA at 50% RH (Equation (6)), which liberated ammonia. It is because of the crystallization of (NH₄)₂SO₄ at <3% RH, which suppressed the displacement of ammonium. When the TEA supply was removed after 24 h, TEAH sulfate decomposed into TEAHBS. The spectrum of TEA-removed reacted NH₄HSO₄ matched the spectrum of the synthesized 50 mol% (NH₄)₂SO₄/TEAHBS mixture at <3% RH (Figure 5b). The EDB-measured mass ratio of the reacted NH₄HSO₄ particle to the initial NH₄HSO₄ is also consistent with the mass ratio of the sum of (NH₄)₂SO₄ and TEAHBS to NH₄HSO₄ at the stoichiometric ratio determined using Equation (8) (Table 1). Unlike TEA reactions of aqueous ammonium salts (e.g., NH₄Cl), the TEA reaction of NH₄HSO₄ under dry condition (<3% RH) did not result in the displacement of ammonium (1667 cm⁻¹). The mass increase of amorphous solid NH₄HSO₄ was attributed to the acid-base neutralization reactions (Qiu et al. 2011) of acidic bisulfate and alkaline TEA (Equation (8)). From the results above, an aqueous phase plays a crucial role in effective uptake of TEA by ammonium salts. Fully deliquesced droplets had much higher TEA uptake compared to those of solid particles. Owing to less severe mass transfer limitation, the amorphous solids tended to have enhanced uptake over those of the crystalline solids.

FIG. 6 Uptake kinetics of aqueous ammonium sulfate particles at 3-ppm and 700-ppb TEA, all at 50% RH. (Color figure available online.)

6. Uptake at Low TEA Concentrations

Low TEA concentration (3-ppm or 700-ppb) uptake experiments were conducted for aqueous (NH₄)₂SO₄ droplets at 50% RH and the kinetics are shown in Figure 6. As in the TEA reaction of (NH₄)₂SO₄ droplet at 40-ppm (50% and 75% RH), TEAH sulfate was formed when the aqueous (NH₄)₂SO₄ droplet (50% RH) was exposed to 700-ppb TEA for 100 h (Supp. Info., Figure S6). The particle that reacted at 700 ppb remained as a droplet and became fluorescent when dried at <3% RH.

For the 20-h 3-ppm TEA uptake experiment using (NH₄)₂SO₄ droplet (50% RH), the displacement of ammonium by TEAH was not fully complete. Hence, the reacted particle was an aqueous mixture of ammonium, TEAH, bisulfate, and sulfate. From the FWHM analysis and the Mie scattering pattern (similar to the analysis of amorphous solid NH₄HSO₄ reaction), crystallization of the unreacted (NH₄)₂SO₄ occurred when the 3-ppm TEA-exposed (NH₄)₂SO₄ particle was dried at <3% RH while the reaction product, TEAHBS, remained as a liquid at <3% RH in the absence of TEA. To determine the ratio of (NH₄)₂SO₄ to TEAHBS, a calibration based on the collected Raman spectra of synthesized mixtures of (NH₄)₂SO₄ and TEAHBS with 11, 20, 43, 69, and 87 mol% (NH₄)₂SO₄ was established (Supp. Info., Figure S7). The normalized sulfate peak area ratio (A_sulfate), defined as the area of sulfate (613 cm⁻¹) over the sum of the areas of sulfate (613 cm⁻¹) and bisulfate (583 cm⁻¹), was plotted against the mol% of (NH₄)₂SO₄ (Figure 7).

FIG. 7 Normalized sulfate peak area ratio against mol% ammonium sulfate. (Color figure available online.)

FIG. 8 Hygroscopicities of E-AIM predictions of ammonium sulfate and EDB-measurements of ammonium sulfate particles, the 3-ppm TEA-exposed (20 h, 50% RH) ammonium sulfate particle and a TEAHBS particle, and Mie scattering patterns of the 3-ppm TEA-exposed ammonium sulfate particle. (Color figure available online.)

The A_sulfate of 3-ppm TEA-exposed (NH₄)₂SO₄ was 0.365. From Figure 7, the corresponding mol% of (NH₄)₂SO₄ was 41% (and the mol% of TEAHBS was 59%). The mass ratio of exposed particle to initial (NH₄)₂SO₄ (on a dry basis) particle, calculated using the mole ratio of (NH₄)₂SO₄ to TEAHBS, was 1.30, consistent with the EDB-measured mass ratio of 1.29.

The hygroscopicity of (NH₄)₂SO₄ particle reacted at 50% RH (in aqueous state) with 3-ppm TEA for 20 h was measured and compared with those of unexposed (NH₄)₂SO₄ particle, TEAHBS particle and E-AIM predictions of (NH₄)₂SO₄ (Figure 8).The EDB-measured hygroscopicity of unexposed (NH₄)₂SO₄ particle matched the E-AIM predictions well (Clegg et al. 1998) and (NH₄)₂SO₄ was observed to deliquesce at about 80% RH. Each TEAHBS ion had 3 extra hydrophobic ethyl groups, making the TEAHBS particle less hygroscopic than (NH₄)₂SO₄. The hygroscopicity of 3-ppm TEA-exposed (NH₄)₂SO₄ particle (determined as a mixture of ∼40 mol% (NH₄)₂SO₄ and ∼60 mol% TEAHBS) was between those of unexposed (NH₄)₂SO₄ particle and TEAHBS particle and showed partial deliquescence and efflorescence properties. The TEA-exposed (NH₄)₂SO₄ particle completely deliquesced at about 70.5% RH (open diamonds), as reflected by the formation of regular fringes. The hygroscopicity of the completely deliquesced droplet was similar to that of unexposed (NH₄)₂SO₄ particle. The (NH₄)₂SO₄ portion of TEA-exposed (NH₄)₂SO₄ crystalized at 47–51% RH (closed diamonds), as supported by the FWHM analysis.

As shown in Figure 8, at <3% RH, the crystalline (NH₄)₂SO₄ particle (unexposed) exhibited an irregular scattering pattern while the TEA-exposed (NH₄)₂SO₄ particle (reacted at 50% RH but dried at <3% RH subsequently) exhibited a pattern of distorted fringes owing to the mixing of liquid TEAHBS and crystalline (NH₄)₂SO_4.

Using a smog chamber, Murphy et al. (2007) studied the uptake of several short-chain amines by submicron (NH₄)₂SO₄ particles at 10% RH. The particle composition was measured simultaneously by an Aerodyne Time of Flight Aerosol Mass Spectrometer (cToF-AMS) and a Particle-Into-Liquid Sampler coupled to Ion Chromatography (PILS-IC). They found a sudden increase in AMS-measured sulfate concentration when the uptake of amines started. However, the sulfate concentration measured by IC remained constant. They speculated that the amine reactions of (NH₄)₂SO₄ may have turned the particle more liquid-like and increased the collection efficiency of the AMS. From this study, we showed the formation of liquid (or supercooled liquid) aminium sulfates (e.g., TEAHBS) as a product of TEA and (NH₄)₂SO₄ reaction. The formation of liquid aminium sulfates may help explain the increase in particle collection efficiency in the AMS/chamber experiments of amine uptake by (NH₄)₂SO₄. The reaction was more favorable in aqueous condition; yet, we still observed a <5% mass increase by crystalline (NH₄)₂SO₄ under<3% RH condition. The mass increase at 10% RH is expected to be larger since TEAHBS uptakes water continuously, according to our measurements. Furthermore, the submicron particles used in Murphy et al. (2007) are significantly smaller than the particles we used here and hence they are expected to have a larger mass transfer rate.

CONCLUSIONS AND ATMOSPHERIC IMPLICATIONS

Using a coupled EDB/Raman system, the heterogenous uptakes of TEA vapor by aqueous and solid (crystalline or amorphous) salts of common inorganic aerosol surrogates were investigated. Although we used particles much larger than atmospheric particles, the effects of the phase state of the particles on uptake of TEA were clearly observed. Aqueous droplets of (NH₄)₂SO₄ and NH₄HSO₄ (50 or 75%) underwent an overwhelming equilibrium mass increase of >90% after exposure to 40-ppm TEA and a complete displacement of ammonium by TEAH was observed. However, crystalline solid (NH₄)₂SO₄ and (NH₄)₂C₂O₄ were comparatively inert (<5% mass increase) to TEA uptake, presumably due to the diffusion limitation in these solids. Aqueous-phase chemistry played a significant role in ammonium displacement. Unlike the crystalline solids, amorphous NH₄NO₃ had a mass increase of >90% after an exposure to 40-ppm TEA at <3% RH. Moreover, amorphous NH₄HSO₄ (<3% RH) was also effectively neutralized by TEA, resulting in an equilibrium mass increase of about 50%. Enhanced TEA uptake by amorphous solid over that of crystalline solid in these experiments was attributed to the reduced mass transfer limitation. Amorphous and liquid-like particles have been postulated in the atmosphere (Mikhailov et al. 2009). Enhanced amine uptake by amorphous particles over those of crystalline particles in the atmosphere is expected. Qiu et al. (2011) studied the reaction of alkylamines with NH₄HSO₄ and proposed the formation of alkylaminium ammonium sulfate based on vapor phase measurements by ID-CIMS. They suggested that NH₄HSO₄ was neutralized by vapors of alkylamines but did not address the bulk composition and the physical states of products. In this study, using Raman spectroscopy, we showed that the reaction of TEA and NH₄HSO₄ at <3% RH produced solid (NH₄)₂SO₄ and liquid TEAH bisulfate/sulfate. We also examined for the first time to our knowledge the inter-conversion between TEAH bisulfate and sulfate.

Recent field observations have suggested that cloud and fog processing (Rehbein et al. 2011) enhanced the gas-to-particle partitioning of trimethylamine into sulfate-containing particles. Ambient acidic sulfate aerosols tend to have a higher fraction of TMA. According to the results of the current study, acidic NH₄HSO₄ can contribute to particulate TEAH formation through both neutralization reaction (bisulfate and TEA) and ammonium displacement. However, for (NH₄)₂SO₄, TEAH formation can only be effective through ammonium displacement in the aqueous phase. Acidic NH₄HSO₄ had a higher potential to react with TEA through neutralization. This may help explain the increased particulate fractions of amines in acidic aerosol. In ambient environments, low RH conditions may induce potential crystallization of (NH₄)₂SO₄ or ammoniated sulfates (Clegg et al. 1998), which in turn would suppress further aqueous ammonium displacement. Fog events and high humidity may favor the formation of particle-phase amines. We noted that the ambient concentration of ammonia is 1–3 orders higher than those of amines. Experiments involving the co-absorption of different amines, and amines/ammonia mixtures are warranted for a more complete understanding of amine uptake.

Acknowledgments

This work was supported by grants from the Research Grants Council of the Hong Kong Special Administrative Region, China (GRF project nos. 600208 and 610909).

[Supplementary materials are available for this article. Go to the publisher's online edition of Aerosol Science and Technology to view the free supplementary files.]